News

Apple buys startup AI Music… so what might its plans be?


Tags:

Apple appears to have acquired UK-based AI Music, a startup founded to build systems for generating music using artificial intelligence, and adapting existing tracks into new forms. Bloomberg reported the acquisition late last night, saying that the deal “was completed in recent weeks”, although Apple declined to comment.

Backup comes from recent company filings by AI Music, with the company’s registered office changed to 100 New Bridge Street in London – also a corporate address used by Apple – and the appointment of Peter Denwood as a director. Denwood being Apple’s director, corporate law, international. Similar filings led us to break the news of Apple buying music analytics startup Semetric back in 2015.

There’s an irony in Apple, which launched its Apple Music streaming service in 2015 with much talk about the importance of humans alongside algorithms, buying a startup that has developed algorithms for creating music.

But of course, beneath that early bluster was some hard work on recommendation algorithms as well as a human curation team, while AI Music has also made a habit of bringing human musicians into its orbit to work with its systems.

The key question concerns what Apple has planned for those systems now. Let’s recap on what AI Music was. We first spotted it early in 2017, a few months before it emerged from stealth mode as a participant in the Abbey Road Red incubator.

The company stressed that “we’re not generating music from scratch… We’re looking at using AI to shift the way in which music is consumed” with a technology for “shape-changing” music from one genre or mood to another.

“It’s that idea of contextual AI. Maybe you listen to a song and in the morning it might be a little bit more of an acoustic version,” CEO Siavash Mahdavi told us in August 2017. “Maybe that same song when you play it as you’re about to go to the gym, it’s a deep-house or drum’n’bass version. And in the evening it’s a bit more jazzy. The song can actually shift itself. The entire genre can change, or the key it’s played in.”

Since those early days, AI Music raised £5m of fundinglaunched a test app called Ossia that enabled people to mix and match vocals and backing tracks; developed ‘Sympaphonic Ads’ technology to match AI-generated music for audio ads with the music being played around them; struck a deal with social network Hornet to create ‘adaptive soundtracks’ for people’s posts; and launched a spin-off called VentureSonic to create music for brands.

Apple Music is a premium-only music streaming service, so in its current form would have little use for AI Music’s ad-tech offshoots – although perhaps it could find a use for them within Apple Podcasts. Meanwhile, AI Music’s original ‘shape-changing’ technology has proved a tough sell for music rightsholders, as another startup trying it, Loudly, explained to Music Ally last May.

AI Music’s team and technology were smart, so Apple could have plenty of other uses for them. This kind of adaptive AI can be put to use to create game soundtracks, so might fit into Apple’s tools for iOS game developers. It might be able to play a role in Apple Fitness+ by adapting music to match the intensity of workouts.

AI Music dabbled in mood-music – a now-deleted YouTube video in 2020 of AI-generated lo-fi hip-hop – so Apple Music could deploy its technology for something in that vein. Or AI Music’s systems could be integrated into Apple’s music-making software like GarageBand and Logic Pro, as assistive technology for human musicians.

This may be the most likely scenario. Back in 2017, Spotify hired one of the key figures in AI-generated music, François Pachet, to (according to his LinkedIn profile) “develop the next generation of AI-assisted music composition tools”. Some of those tools have since been used for the ‘American Folk Songs’ project by artist Benoit Carré, under his Skygge pseudonym.

AI Music could bring Apple similar capabilities – or at least help any ambitions to develop them beyond the automated features that are already in its music software.

So, while today’s news may spark some fears about two of the world’s biggest music-streaming services having AI-generated music capabilities in-house, the real impact may be on helping humans to make music in new ways, rather than replacing those humans with AIs.


Written by: Stuart Dredge